Open Forum Infectious Diseases MAJOR ARTICLE

Open Forum Infectious Diseases MAJOR ARTICLE Comparative Effectiveness of Prophylactic Strategies for Perinatal Transmission of Hepatitis B Virus: A Network Meta-analysis of Randomized Controlled Trials Zhi-Xian Chen, 1,* Xun Zhuang, 2,* Xiao-Hong Zhu, 3 Yan-Li Hao, 3 Gui-Fang Gu, 4 Meng-Zhi Cai, 3 and Gang Qin 3 1 Department of Clinical Pharmacy, Nantong Health College of Jiangsu Province, Jiangsu, China; 2 Department of Epidemiology and Biostatistics, School of Public Health, Nantong University, Jiangsu, China; 3 Center for Liver Diseases and 4 Department of Obstetrics and Gynecology, Nantong Third People s Hospital affiliated to Nantong University, Jiangsu, China Background. Perinatal transmission is the main route of hepatitis B virus (HBV) transmission. While several measures have been attempted as means of preventing perinatal HBV transmission, the optimal strategy remains inconclusive. Methods. We conducted a comprehensive search, through December 2016, for randomized controlled trials (RCTs) that compared the following measures among pregnant women with HBV infection: placebo/none, active immunoprophylaxis (hepatitis B vaccine series starting at birth [HBVac]), passive-active immunoprophylaxis (hepatitis B immunoglobulin and vaccine [HBIG+HBVac]), prenatal HBIG administration (HBIG/HBIG+HBVac), and prenatal antiviral therapy (AVT/HBIG+HBVac). Direct, indirect, and network meta-analyses were performed for all treatment comparisons. Results. Fifteen RCTs involving 2706 infants of HBV carrier mothers were eligible for analysis. Network meta-analysis demonstrated similar results as direct and indirect comparisons. HBVac alone significantly reduced the risk of hepatitis B infection in infants of HBV carrier mothers (relative risk [RR], 0.32; 95% confidence interval [CI], 0.21 0.50). The combination of immunoglobulin with vaccine is superior to vaccine alone (RR, 0.37; 95% CI, 0.20 0.67). Prenatal HBIG administration and antiviral therapy offer further advantages over current passive-active immunoprophylaxis for infants of highly viremic (HBV DNA level higher than 2 10 5 IU/ ml) mothers (RR, 0.47; 95% CI, 0.29 0.75; and RR, 0.31; 95% CI, 0.10 0.99, respectively). There was no significant publication bias. Conclusions. Based on the universal infantile vaccination program, HBIG for infants born to HBV carrier mothers further reduces transmission. For highly viremic mothers whose children are still at risk for transmission under current immunoprophylaxis, prenatal HBIG administration or antiviral therapy in late pregnancy may be considered if more long-term evidence supports its efficacy and safety. Keywords. hepatitis B virus; network meta-analysis; perinatal transmission; prophylaxis. Chronic hepatitis B virus (HBV) infection is a major cause of liver cirrhosis and cancer globally, attributing a heavy disease burden. More than 90% of perinatal infected newborns become chronic carriers, while infected children and adolescents/adults have lower risks of chronicity (20% 40% and 0% 10%, respectively) [1]. Perinatal transmission contributes significantly to the high HBV prevalence in endemic countries. Since the introduction of hepatitis B vaccine (HBVac), active immunoprophylaxis has proven efficient for preventing perinatal HBV transmission. The World Health Organization (WHO) recommends that all countries include HBVac starting with a Received 21 August 2017; editorial decision 9 October 2017; accepted 10 October 2017. * Equal contribution. Correspondence: G. Qin, MD, PhD, Center for Liver Diseases, Nantong Third People s Hospital affiliated to Nantong University, 60 Mid-Youth Road, Nantong 226006, China (tonygqin@ntu.edu.cn). Open Forum Infectious Diseases The Author 2017. Published by Oxford University Press on behalf of Infectious Diseases Society of America. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (http://creativecommons.org/licenses/ by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contact journals.permissions@oup.com DOI: 10.1093/ofid/ofx225 birth dose into routine national infant immunization programs [2]. However, for infants of HBV carrier (hepatitis B surface antigen/hbsag positive) mothers, even after being vaccinated with both hepatitis B immunoglobulin and vaccine (HBIG+HBVac, termed passive-active immunoprophylaxis), 10% 20% of them may still become chronic HBV carriers, mostly due to intrauterine infection [3]. Several studies reporting effects for maternal HBIG administration or antiviral therapy (AVT) during late pregnancy have not yet provided conclusive evidence. Most of these studies might have potentially biased results due to lack of randomization. There is paucity of head-to-head randomized controlled trials (RCTs) comparing different prophylactic interventions, which can inform clinicians regarding the comparative effectiveness of these interventions. Moreover, direct pairwise meta-analyses can only answer questions about pairs of treatments with partial information. In contrast, network meta-analyses combine the simultaneous analyses of direct and indirect evidence for mixed-effect estimate calculation, hence helping to evaluate the comparative effectiveness of multiple interventions [4]. In this study, we performed direct, indirect, and network meta-analyses to assess the relative efficacy of all prophylactic Perinatal HBV Transmission Prophylaxis OFID 1

interventions (HBVac, HBIG+HBVac, HBIG/HBIG+HBVac, and AVT/HBIG+HBVac) for the management of HBV carrier mothers and their infants. METHODS Search Strategy We included the following 4 interventions: (1) infants were injected with a 3- or 4-dose vaccine series starting within 24 hours after delivery (HBVac); (2) infants were injected with both HBIG and HBVac (HBIG+HBVac); (3) pregnant women were injected with multiple doses of HBIG in the third trimester of pregnancy (typically on weeks 28, 32, and 36) while infants were injected with HBIG and HBVac (HBIG/HBIG+HBVac); (4) pregnant women with a high viral load (HBV DNA level higher than 2 10 5 IU/mL) were given antiviral therapy (lamivudine or telbivudine or tenofovir) during late pregnancy while infants were injected with HBIG and HBVac (AVT/HBIG+HBVac). Eligible studies were predefined as RCTs comparing any intervention with placebo/none or another intervention for HBV carrier mothers and their infants. PubMed, Ovid, Embase, and Cochrane Library were searched from the date of their inception to December 31, 2016, using combinations of the following terms: HBV (or hepatitis B virus ) and perinatal transmission (or mother-to-child transmission or MTCT or intrauterine infection or ectopic or pregnant or pregnancy or mother or children or infant or newborn ) and prophylaxis (or hepatitis B vaccine or HB vaccine or hepatitis B immunoglobulin or HBIG or lamivudine or telbivudine or tenofovir ). Data Extraction and Quality Assessment Two investigators (Z-XC and Y-LH) independently selected studies, reviewed the reports, and recorded information using a predefined data extraction sheet that included relevant items: the first author s name, year of publication, region, study design, inclusion/exclusion criteria, sample size, duration of follow-up, and outcome data. Perinatal HBV transmission, the primary outcome, was defined as seropositivity rate for HBsAg or HBV DNA in the infants both at birth and at 6 12 months of follow-up. Two investigators (G-FG and M-ZC) independently assessed the risk of bias and the quality of studies according to the Jadad score [5]. Discrepancies were resolved by consensus and consultation with other investigators (XZ and GQ). Statistical Analysis Data analysis was performed according to the per-protocol principle. All statistical analyses were conducted using Stata version 14.0 (StataCorp, TX). Direct meta-analysis with the random-effects model was applied to estimate pooled relative risks (RRs) and 95% confidence intervals (CIs). Indirect comparisons were performed using the adjusted indirect method described by Miladinovic et al. [6]. The network meta-analysis models, network diagram, predictive interval (PrI) plot, and ranking plots were constructed using commands of the network package in Stata [7]. In network meta-analysis, the RRs were presented with 95% CIs alongside their 95% PrIs to provide intervals within which the estimates of future studies are expected to be. The heterogeneity test was conducted with the chi-square test and I 2. An I 2 index of 50% or more indicated a high degree of heterogeneity. Publication bias was assessed by examining Begg s funnel plots and Egger s regression tests. We estimated number needed to treat (NNT) from the summary estimates for the primary outcome using the calculated absolute perinatal transmission risk with HBVac alone. 1524 records identified through search strategy 1165 excluded by screening titles/abstracts HBVac 9 (n = 746) 3 (n = 247) 359 full texts retrieved for detail evaluation HBIG + HBVac 3 (n = 216) Placebo/none 43 records added by cross referencing Figure 1. trial. 15 RCTs included in the meta-analysis Excluded due to: 342 nonrandomized studies 27 duplicate published studies 18 no outcome of interest Flowchart of study selection. Abbreviation: RCT, randomized controlled 3 (n = 1517) 3 (n = 325) HBIG/HBIG + HBVac AVT/HBIG + HBVac Figure 2. Network plot for different prophylactic interventions to prevent perinatal transmission among infants of hepatitis B virus carrier mothers. Abbreviations: AVT, antiviral therapy; HBIG, hepatitis B immunoglobulin; HBVac, hepatitis B vaccine. 2 OFID Chen et al

RESULTS Characteristics and Quality of Included Studies We identified 1524 citations through our searches, from which 15 clinical trials involving 2706 infants of HBV carrier mothers were selected (Figure 1). Some studies had more than 2 arms, resulting in up to 21 relevant comparisons: HBVac vs placebo/none: 3 trials [8 10]; HBIG+HBVac vs placebo/none: 3 trials [8 10]; HBIG+HBVac vs HBVac: 9 trials [8 16]; HBIG/HBIG+HBVac vs HBIG+HBVac: 3 trials [17 19]; AVT+HBIG+HBVac vs HBIG+HBVac: 3 trials [20 22]. Five RCTs listed in the meta-analysis of Brown et al. [23] were not included in our study because their randomization was questioned [24]. Figure 2 demonstrates all available direct comparisons and the network. The mean sample size was 188 infants, ranging between 39 and 987. Fourteen (93.3%) trials recruited patients from Asia (mainland China: 7 trials; Thailand: 3 trials; India: 2 trials; Hong Kong: 1 trial; Taiwan: 1 trial), and 1 (6.7%) recruited patients from Oceania (New Zealand). Thirteen were published in English, and 2 were published in Chinese. Eight studies only recruited HBsAg- and HBeAg-positive mothers [8, 9, 11 13, 15, 18, 22]. Five studies had clear classifications of HBeAg-positive and HBeAg-negative mothers, with comparable numbers in experimental and control groups [14, 16, 17, 19, 20]. Two studies did not report the HBeAg status of enrolled HBV carrier mothers [21]. Eight studies reported the method to generate the sequence of randomization [8, 10, 15 17, 19, 20, 22]. Four studies adopted placebo and described the method of double blinding [8, 10, 16, 20]. Nine studies recorded withdrawals and dropouts. Studylevel quality assessments with Jadad scores are summarized in Table 1. Direct and Indirect Meta-Analyses Three studies compared the efficacy of hepatitis B vaccine with placebo or no treatment among infants of HBV carrier mothers. At age 6 months or older, the infant HBsAg-positive rates were 20.2% (26/129) in the HBVac group and 59.3% (70/118) in the placebo/none group (RR, 0.33; 95% CI, 0.23 0.48; I 2, 22.7%). Nine studies comparing HBIG+HBVac vs HBVac demonstrated perinatal transmission rates of 3.6% (13/359) in the HBIG+HBVac group and 11.6% (45/387) in the HBVac group (RR, 0.34; 95% CI, 0.19 0.60; I 2, 0%). For infants of highly viremic mothers, both HBIG/HBIG+HBVac (3 RCTs) and ATV+HBIG+HBVac (3 RCTs) led to further lower perinatal transmission rates compared with HBIG+HBVac. At age of 6 months or more, the infant HBsAg-positive rates were 5.1% (39/768) in the HBIG/HBIG+HBVac group and 11.2% (84/749) in the HBIG+HBVac group (RR, 0.46; 95% CI, 0.32 0.65; I 2, 59.9%). Likewise, the perinatal transmission rates were 1.8% (3/163) in the AVT/HBIG+HBVac group and 9.9% (16/162) in the HBIG+HBVac group (RR, 0.24; 95% CI, 0.09 0.69; I 2, 18.3%). For the above direct comparisons, the results of indirect meta-analyses were largely similar, with overlapping confidence intervals. In addition, indirect meta-analyses demonstrated that HBIG/HBIG+HBVac was superior to placebo/none and HBVac in reducing perinatal transmission (RR, 0.05; 95% CI, 0.02 0.13; and RR, 0.15; 95% CI, 0.05 0.42, respectively). ATV/ HBIG+HBVac also further reduced transmission, as compared with placebo/none and HBVac (RR, 0.03; 95% CI, 0.01 0.13; and RR, 0.10; 95% CI, 0.02 0.42, respectively). AVT/HBIG+HBVac and HBIG/HBIG+HBVac appear comparable with each other for decreasing perinatal transmission risk (RR, 0.63; 95% CI, 0.16 2.48). Network Meta-analyses Compared with HBVac alone, HBIG+HBVac (RR, 0.37; 95% CI, 0.20 0.67), HBIG/HBIG+HBVac (RR, 0.17; 95% CI, 0.08 0.37), and AVT/HBIG+HBVac (RR, 0.11; 95% CI, 0.03 0.42) further reduced perinatal HBV transmission (Figure 3). Using the absolute transmission risk of 11.6 (45/387) with HBVac, the corresponding NNTs for HBIG+Vac, HBIG/HBIG+Vac, and AVT/HBIG+Vac to prevent 1 perinatal infection were 14, 11 and 10, respectively. All the above 3 interventions were superior to HBVac in reducing perinatal HBV transmission. For infants of highly viremic mothers, both HBIG/ HBIG+HBVac and AVT/HBIG+HBVac were superior to HBIG+HBVac in reducing perinatal infection (RR, 0.47; 95% CI, 0.29 0.75; and RR, 0.31; 95% CI, 0.10 0.99, respectively) (Figure 3). The combination of prenatal antiviral therapy with passive-active immunoprophylaxis had the highest probability of being ranked best (for decreasing perinatal transmission), and the combination of prenatal HBIG administration with passive-active immunoprophylaxis had a high probability of being ranked second (Figure 4). Safety Profiles Both hepatitis vaccine and immunoglobulin appeared to be safe, while a few studies in this meta-analysis reported adverse effects without mentioning details. Data on maternal and fetal safety with antivirals remain limited. Publication Bias In general, there was no evidence of publication bias, both qualitatively based on funnel plot symmetry (Figure S1) and quantitatively based on Egger s regression tests (P >.05 for all comparisons). No significant differences were found between direct and indirect estimates where both were available. Besides, the 3 methods had overlapping CIs for all interventions (Table 2). DISCUSSION To the best of our knowledge, this is the first study that has combined both direct and indirect evidence in terms of comparative Perinatal HBV Transmission Prophylaxis OFID 3

Table 1. Characteristics of Included RCTs Study Region Period Center Mother HBeAg+/- Prophylaxis Strategy Mother Child Baby Sample Size No. of Transmissions Jadad Score Wong VC (1984) Hong Kong 1981 1983 1 All positive C: none C: Placebo C: 34 C: 25 5 (2 + 2 + 1) T1: none T1: HBVac T1: 35 T1: 7 T2: none T2: HBIG+HBVac T2: 35 T2: 2 Lo KJ (1985) Taiwan 1981 1983 1 All positive C: none C: none C: 29 C: 26 2 (1 + 0 + 1) T1: none T1: HBVac T1: 38 T1: 9 T2: none T2: HBIG+HBVac T2: 36 T2: 4 Farmer K (1987) New Zealand 1983 1985 1 All positive C: none C: HBVac C: 18 C: 4 2 (1 + 0 + 1) T: none T: HBIG+HBVac T: 21 T: 3 Theppisai U (1987) Thailand 1984 1985 1 All positive T1: none T1: HBVac T1: 18 T1: 2 1 (1 + 0 + 0) T2: none T2: HBIG+HBVac T2: 27 T2: 0 Poovorawan Y (1992) Thailand NR 1 All positive C: none C: HBVac C: 59 C: 2 2 (1 + 0 + 1) T: none T: HBIG+HBVac T: 60 T: 0 Sehgal A (1992) India 1987 1989 1 T1: 7/14 T1: none T1: HBVac T1: 21 T1: 2 2 (1 + 0 + 1) T2: 7/17 T2: none T2: HBIG+HBVac T2: 24 T2: 2 Assateerawatt A (1993) Thailand 1988 1989 1 All positive T1: none T1: HBVac T1: 26 T1: 3 3 (2 + 0 + 1) T2: none T2: HBIG+HBVac T2: 23 T2: 1 Xu ZY (1995) China 1982 1989 1 NR C: none C: placebo C: 55 C: 19 4 (2 + 2 + 0) T1: none T1: HBVac T1: 56 T1: 10 T2: none T2: HBIG+HBVac T2: 27 T2: 1 Zhu QR (2003) China 1995 1999 1 C: 189/304 C: none C: HBIG+HBVac C: 496 C: 48 2 (2 + 0 + 0) T: 169/318 T: HBIG 200 400U (7,8,9m) T: HBIG+HBVac T: 491 T: 19 Yuan J (2006) China 1999 2004 1 All positive C: none C: HBIG+HBVac C: 133 C: 17 2 (1 + 0 + 1) T: HBIG 200 400U (7,8,9m) T: HBIG+HBVac T: 118 T: 13 Wang FY (2008) China 2001 2006 1 C: 60/60 C: none C: HBIG+HBVac C: 120 C: 19 2 (2 + 0 + 0) T: 79/80 T: HBIG 200 400U (7,8,9m) T: HBIG+HBVac T: 159 T: 7 Xu WM (2009) China NR 7 C: 61/0 C: placebo C: HBIG+HBVac C: 44 C: 6 5 (2 + 2 + 1) T: 88/1 T: LAM 100mg/d from 32w of gestation to 1m after delivery T: HBIG+HBVac T: 40 T: 3 Zhang LJ (2009) China 2007 2008 1 NR C: none T1: HBIG+HBVac C: 30 C: 4 1 (1 + 0 + 0) T: LdT 600mg/d from 28 32w of gestation to 1m after delivery T2: HBIG+HBVac T: 31 T: 0 Pande C (2013) India 2004 2009 1 19% positive C: none C: Placebo+HBVac C: 116 C: 6 5 (2 + 2 + 1) T: none T: HBIG+HBVac T: 106 T: 0 Pan CQ (2016) China 2012 2013 6 All positive C: none C: HBIG+HBVac C: 88 C: 6 3 (2 + 0 + 1) T: TDF 300mg/d from 30 32w of gestation to 1m after delivery T: HBIG+HBVac T: 92 T: 0 Abbreviations: C, control; HBIG, hepatitis B immunoglobulin; HBVac, HBV vaccine; Jadad score, consists of randomization (2 points) + blindness (2 points) + withdrawal (1 point); LAM, lamivudine; LdT, telbivudine; NR, not reported; RCT, randomized controlled trial; T, treatment; TDF, tenofovir disoproxil fumarate. 4 OFID Chen et al

Prophylactic Strategy RR with 95% CI and 95%Prl HBVac vs placebo/none HBIG + HBVac vs placebo/none HBIG/HBIG + HBVac vs placebo/none AVT/HBIG + HBVac vs placebo/none HBIG + HBVac vs HBVac HBIG/HBIG + HBVac vs HBVac AVT/HBIG + HBVac vs HBVac HBIG/HBIG + HBVac vs HBIG + HBVac AVT/HBIG + HBVac vs HBIG + HBVac AVT/HBIG + HBVac vs HBIG/HBIG + HBVac 0.32 (0.21 0.50) (0.15 0.68) 0.12 (0.06 0.22) (0.05 0.29) 0.06 (0.03 0.12) (0.02 0.16) 0.04 (0.01 0.14) (0.01 0.18) 0.37 (0.20 0.67) (0.15 0.89) 0.17 (0.08 0.37) (0.06 0.48) 0.11 (0.03 0.42) (0.02 0.57) 0.47 (0.29 0.75) (0.22 1.01) 0.31 (0.10 0.99) (0.07 1.32) 0.66 (0.19 2.31) (0.14 3.09) 0 0.7 1 3.3 Figure 3. Predictive interval plot for the network of various prophylactic interventions for perinatal hepatitis B virus transmission. Abbreviations: AVT, antiviral therapy; HBIG, hepatitis B immunoglobulin; HBVac, hepatitis B vaccine. effectiveness of available prophylaxis measures for the prevention of perinatal HBV transmission. Based on the analysis of the 15 currently available RCTs involving 2706 infants of HBV carrier mothers, we made several key observations to estimate the relative efficacy of all prophylactic interventions to interrupt perinatal HBV transmission: (1) Hepatitis B vaccine alone significantly 100 Placebo/none HBVac HBIG + HBVac 50 Probability, % 0 100 HBIG/HBIG+HBVac AVT/HBIG + HBVac Best 2nd 3rd 4th Worst Rank 50 0 Best 2nd 3rd 4th Worst Best 2nd 3rd Rank Rank 4th Worst Figure 4. Rankograms for decreasing perinatal hepatitis B virus transmission risk and the surface under the cumulative ranking curve (SUCRA) for each prophylactic intervention. Abbreviations: AVT, antiviral therapy; HBIG, hepatitis B immunoglobulin; HBVac, hepatitis B vaccine. Perinatal HBV Transmission Prophylaxis OFID 5

Table 2. Pooled Summary Estimates of Evidence Derived From Direct or Indirect Meta-analysis Informing on Comparative Effectiveness of Prophylaxis Strategies for Reducing Perinatal Transmission of HBV Direct Estimates Indirect Estimates Prophylaxis Strategies RR (95% CI) Events, n/total, N I 2, % RR (95% CI) Compared with placebo/none HBVac 0.33 (0.23 0.48) 26/129 vs 70/118 22.7% 0.28 (0.11 0.70) HBIG+HBVac 0.10 (0.05 0.22) 7/98 vs 70/118 0 0.13 (0.06 0.26) HBIG/HBIG+HBVac 0.05 (0.02 0.13) AVT/HBIG+HBVac 0.03 (0.01 0.13) Compared with HBVac HBIG+HBVac 0.34 (0.19 0.60) 13/359 vs 45/387 0 0.32 (0.14 0.75) HBIG/HBIG+HBVac 0.15 (0.05 0.42) AVT/HBIG+HBVac 0.10 (0.02 0.42) Compared with HBIG+HBVac HBIG/HBIG+HBVac 0.46 (0.32 0.65) 39/768 vs 84/749 59.9 AVT/HBIG+HBVac 0.24 (0.09 0.69) 3/163 vs 16/162 18.3 Compared with HBIG/HBIG+HBVac AVT/HBIG+HBVac 0.63 (0.16 2.48) Abbreviations: AVT, antiviral therapy; CI, confidence interval; HBIG, hepatitis B immunoglobulin; HBVac, HBV vaccine; RR, relative risk. decreased the risk of HBV infection among infants of HBV carrier mothers. (2) The combination of immunoglobulin with vaccine is superior to vaccine alone. (3) Maternal HBIG administration and antiviral therapy offer further advantages over passive-active immunoprophylaxis for highly viremic mothers. Antibodies to HBsAg (Anti-HBs) are produced in response to exposure to HBV envelope protein HBsAg and confer protective immunity. Plasma-derived HBVac was first approved in 1981, while genetically engineered recombinant HBVac has been implemented since 1986 and is the current vaccine used worldwide [25]. Implementation of a universal HBVac program has proven successful in preventing infection and related complications. In our network meta-analysis, HBVac alone significantly reduced perinatal transmission risk by nearly 70%. However, in 2006, the reported proportion of newborn infants who received a birth dose of HBVac was only 27% globally, and it was 36% for the 87 countries with high endemicity of chronic HBV infection [26]. Efforts are still needed to improve global coverage of hepatitis B vaccination as well as the other strategies to prevent perinatal transmission for infants of HBV carrier mothers. HBIG, prepared from the plasma of donors with high concentrations of anti-hbs, provides temporary protection (3 6 months) and is usually implemented as an adjunct to the hepatitis B vaccine for postexposure prophylaxis. While the response of neonates to the HBVac is satisfactory, it takes time to develop protective antibodies after vaccination. HBIG is offered to neonates of HBV carrier mothers in order to reduce the time window between exposure to infected mothers and production of anti-hbs induced by vaccine. Here, the combined preventive efficacy was 88%, compared with 68% for vaccine alone. It has been well accepted that HBIG should be given as soon as possible after birth and that it does not need to be repeated if the infant is subsequently vaccinated. The effect of maternal HBIG administration, in terms of prevention of HBV transmission to infants, has been evaluated in quite a few studies in China. Our study found that maternal HBIG administration in the third trimester could further reduce the transmission risk by half, compared with infantile passive-active immunoprophylaxis. This result is in accordance with the previous published meta-analyses [27 29]. Nevertheless, most of the data in previous studies were derived from nonrandomized studies, which were subject to significant biases, especially selection bias. In contrast, we excluded those studies that claimed RCTs but where the randomization was questioned [24]. Thus, all included studies had low risk of bias. The possible mechanism for maternal HBIG administration might be that it binds HBsAg and reduces HBV levels in pregnant women, or transports HBsAg across the placenta and facilitates humoral immunity [27]. More evidence is still needed to confirm these assumptions. Maternal viral load, rather than HBeAg status, has been extensively documented as the most important risk factor for perinatal HBV transmission [30]. Increasing evidence has shown promising efficacy and safety profiles with the use of nucleoside/nucleotide analogues (NAs) in highly infectious mothers. Current evidence shows that maternal antiviral therapy (lamivudine, telbivudine, or tenofovir) further reduces the transmission risk by nearly 70%, compared with the infantile passive-active immunoprophylaxis. In non-head-to-head studies, tenofovir and telbivudine showed higher rates of HBV DNA suppression, alanine aminotransferase (ALT) normalization, and HBeAg seroconversion than lamivudine [31]. For concerns regarding fetal exposure to these drugs used in pregnancy, no 6 OFID Chen et al

significant safety issues for maternal or fetal outcomes were identified in our meta-analysis of included studies. Moreover, data from the Antiretroviral Pregnancy Registry (available from http://www.apregistry.com/interimreport.aspx) have suggested a favorable safety profile for antiviral therapy in pregnancy. Additionally, cessation of therapy 1 3 months after delivery is recommended in mothers without hepatic flares and without preexisting cirrhosis [32]. There are also concerns that the modes of delivery and feeding might influence the risk of perinatal HBV transmission [33, 34]. Until now, no randomized trials have yet been conducted for comparison of the elective cesarean section vs vaginal delivery, or breastfeeding vs bottle-feeding. Given the evidence from observational studies, both vaginal delivery and breastfeeding are considered safe and low risk for HBV transmission, especially when passive-active immunoprophylaxis is appropriately administered [35]. Admittedly, our study has several limitations. First, given the small number of included studies, small number of studies per comparison, and small number of subjects (<30 subjects per arm) in 5 of the 15 RCTs, small study effects could not be avoided. In terms of probabilistic analysis by relative ranking probabilities, the 2 best options for reducing perinatal transmission were AVT/HBIG+HBVac and HBIG/HBIG+HBVac. There is a high degree of uncertainty about which measure works better. Second, we could not carry out subgroup analyses according to HBeAg status of the mothers, vaccination schedules, types and doses of HB vaccine, doses of HBIG, or types of antiviral agents. As most trials only enrolled mothers positive for both HBsAg and HBeAg, the applicability of our findings to HBeAg-negative mothers is limited. The trials that did not report HBeAg status of mothers have not been excluded from the meta-analysis because this might have led to exclusion of some good quality trials. Recombinant vaccine and plasma-derived vaccine have shown no significant difference in decreasing perinatal transmission [36]. Lamivudine, telbivudine, and tenofovir have been used in the nonpregnant population as potent agents against HBV while only tenofovir has a favorable resistance profile. In the setting of short-term use during late pregnancy, resistance is rare. With only 1 RCT for each agent, we feel it acceptable to combine these trials as an antiviral therapy group. Third, HBIG was used to prevent HBV transmission in infants of HBV carrier mothers even before the availability of the HBVac. Yet we did not include it as an independent measure because it has been replaced by the current passive-active immunoprophylaxis. Last, more long-term safety data and continued epidemiological surveillance of children born to HBV carrier mothers are needed. Multiple steps are involved in preventing perinatal HBV transmission. It is our firm belief that the transmission is preventable in most situations. Ensuring strategies for prevention should be made a priority. Supplementary Data Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author. Acknowledgements Financial support. This work was supported in part by grant number BE2015655 from the Jiangsu Provincial Department of Science and Technology, China, by the Qing Lan Project from Jiangsu Provincial Department of Education, China, by grant number 81370520 from the National Natural Science Foundation of China (NSFC), and by grant number HS2016002 from the Nantong Municipal Bureau of Science and Technology, Jiangsu Province, China. Disclosure. The funders of this study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all data in this study and had final responsibility for the decision to submit for publication. Potential conflicts of interest. All authors: no reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. References 1. McMahon BJ, Alward WL, Hall DB, et al. Acute hepatitis B virus infection: relation of age to the clinical expression of disease and subsequent development of the carrier state. J Infect Dis 1985; 151:599 603. 2. World Health Organization. Hepatitis B vaccines: WHO position paper, July 2017 recommendations. Vaccine 2017; doi: 10.1016/j.vaccine.2017.07.046. 3. Zou H, Chen Y, Duan Z, et al. Virologic factors associated with failure to passive-active immunoprophylaxis in infants born to HBsAg-positive mothers. J Viral Hepat 2012; 19:e18 25. 4. Kiefer C, Sturtz S, Bender R. Indirect comparisons and network meta-analyses. Dtsch Arztebl Int 2015; 112:803 8. 5. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996; 17:1 12. 6. Miladinovic B, Chaimani A, Hozo I, Djulbegovic B. Indirect treatment comparison. Stata J 2014; 14:76 86. 7. White IR. Network meta-analysis. Stata J 2015; 15:951 85. 8. Wong VC, Ip HM, Reesink HW, et al. Prevention of the HBsAg carrier state in newborn infants of mothers who are chronic carriers of HBsAg and HBeAg by administration of hepatitis-b vaccine and hepatitis-b immunoglobulin. Doubleblind randomised placebo-controlled study. Lancet 1984; 1:921 6. 9. Lo KJ, Tsai YT, Lee SD, et al. Combined passive and active immunization for interruption of perinatal transmission of hepatitis B virus in Taiwan. Hepatogastroenterology 1985; 32:65 8. 10. Xu ZY, Duan SC, Margolis HS, et al. Long-term efficacy of active postexposure immunization of infants for prevention of hepatitis B virus infection. United States-People s Republic of China Study Group on Hepatitis B. J Infect Dis 1995; 171:54 60. 11. Farmer K, Gunn T, Woodfield DG. A combination of hepatitis B vaccine and immunoglobulin does not protect all infants born to hepatitis B e antigen positive mothers. N Z Med J 1987; 100:412 4. 12. Theppisai U, Thanuntaseth C, Chiewsilp P, Siripoonya P. A comparison between the efficacy of passive-active and active immunization for prevention of perinatal transmission of hepatitis B virus. J Med Assoc Thai 1987; 70:459 62. 13. Poovorawan Y, Sanpavat S, Pongpunglert W, et al. Long term efficacy of hepatitis B vaccine in infants born to hepatitis B e antigen-positive mothers. Pediatr Infect Dis J 1992; 11:816 21. 14. Sehgal A, Sehgal R, Gupta I, et al. Use of hepatitis B vaccine alone or in combination with hepatitis B immunoglobulin for immunoprophylaxis of perinatal hepatitis B infection. J Trop Pediatr 1992; 38:247 51. 15. Assateerawatt A, Tanphaichitr VS, Suvatte V, Yodthong S. Immunogenicity and efficacy of a recombinant DNA hepatitis B vaccine, GenHevac B Pasteur in high risk neonates, school children and healthy adults. Asian Pac J Allergy Immunol 1993; 11:85 91. 16. Pande C, Sarin SK, Patra S, et al. Hepatitis B vaccination with or without hepatitis B immunoglobulin at birth to babies born of HBsAg-positive mothers prevents overt HBV transmission but may not prevent occult HBV infection in babies: a randomized controlled trial. J Viral Hepat 2013; 20:801 10. Perinatal HBV Transmission Prophylaxis OFID 7

17. Zhu Q, Yu G, Yu H, et al. A randomized control trial on interruption of HBV transmission in uterus. Chin Med J (Engl) 2003; 116:685 7. 18. Yuan J, Lin J, Xu A, et al. Antepartum immunoprophylaxis of three doses of hepatitis B immunoglobulin is not effective: a single-centre randomized study. J Viral Hepat 2006; 13:597 604. 19. Wang FY, Lin P, Zhang HZ. A randomized controlled trial on effect of hepatitis B immune globulin in preventing hepatitis B virus transmission from mothers to infants. Zhonghua Er Ke Za Zhi 2008; 46:61 3. 20. Xu WM, Cui YT, Wang L, et al. Lamivudine in late pregnancy to prevent perinatal transmission of hepatitis B virus infection: a multicentre, randomized, double-blind, placebo-controlled study. J Viral Hepat 2009; 16:94 103. 21. Zhang LJ, Wang L. Blocking intrauterine infection by telbivudine in pregnant chronic hepatitis B patients. Zhonghua Gan Zang Bing Za Zhi 2009; 17:561 3. 22. Pan CQ, Duan Z, Dai E, et al; China Study Group for the Mother-to-Child Transmission of Hepatitis B. Tenofovir to prevent hepatitis B transmission in mothers with high viral load. N Engl J Med 2016; 374:2324 34. 23. Brown RS Jr, McMahon BJ, Lok AS, et al. Antiviral therapy in chronic hepatitis B viral infection during pregnancy: a systematic review and meta-analysis. Hepatology 2016; 63:319 33. 24. Zhou YH. Prevention of mother-to-child transmission of hepatitis B virus by treating mothers with high viral loads. Hepatology 2016; 64:1823 4. 25. Visvanathan K, Dusheiko G, Giles M, et al. Managing HBV in pregnancy. Prevention, prophylaxis, treatment and follow-up: position paper produced by Australian, UK and New Zealand key opinion leaders. Gut 2016; 65:340 50. 26. Worldwide implementation of hepatitis B vaccination of newborns, 2006. Wkly Epidemiol Rec 2008; 83:429 34. 27. Shi Z, Li X, Ma L, Yang Y. Hepatitis B immunoglobulin injection in pregnancy to interrupt hepatitis B virus mother-to-child transmission-a meta-analysis. Int J Infect Dis 2010; 14:e622 34. 28. Xu H, Zeng T, Liu JY, et al. Measures to reduce mother-to-child transmission of Hepatitis B virus in China: a meta-analysis. Dig Dis Sci 2014; 59:242 58. 29. Jin H, Zhao Y, Tan Z, et al. Immunization interventions to interrupt hepatitis B virus mother-to-child transmission: a meta-analysis of randomized controlled trials. BMC Pediatr 2014; 14:307. 30. Chen ZX, Gu GF, Bian ZL, et al. Clinical course and perinatal transmission of chronic hepatitis B during pregnancy: a real-world prospective cohort study. J Infect 2017; 75:146 54. 31. Govan L, Wu O, Xin Y, et al. Comparative effectiveness of antiviral treatment for hepatitis B: a systematic review and Bayesian network meta-analysis. Eur J Gastroenterol Hepatol 2015; 27:882 94. 32. Terrault NA, Bzowej NH, Chang KM, et al; American Association for the Study of Liver Diseases. AASLD guidelines for treatment of chronic hepatitis B. Hepatology 2016; 63:261 83. 33. Pan CQ, Zou HB, Chen Y, et al. Cesarean section reduces perinatal transmission of hepatitis B virus infection from hepatitis B surface antigen-positive women to their infants. Clin Gastroenterol Hepatol 2013; 11:1349 55. 34. Hu Y, Dai X, Zhou YH, Yang H. A knowledge survey of obstetrics and gynecology staff on the prevention of mother-to-child transmission of hepatitis B virus. J Infect Dev Ctries 2013; 7:391 7. 35. Buchanan C, Tran TT. Management of chronic hepatitis B in pregnancy. Clin Liver Dis 2010; 14:495 504. 36. Barbosa C, Smith EA, Hoerger TJ, et al. Cost-effectiveness analysis of the national Perinatal Hepatitis B Prevention Program. Pediatrics 2014; 133:243 53. 8 OFID Chen et al